An array format for biological and chemical analysis holds the promise to rapidly provide accurate results while minimizing labor. [Nature Genetics, 1999 Vol. 21 (1) supplement pp. 3-4]. Typically, arrays of biological probes such as DNA, RNA or protein molecules are formed either by deposition and immobilization or by in-situ synthesis on inert substrates. In these prior art methods, array formation is usually accomplished by attaching probe molecules directly to a substrate, which may be composed of organic materials (such as polymeric materials like nitrocellulose) or inorganic materials (such as glass or silicon).
The use of silicon as a substrate provides certain advantages related to the well-established methods of semiconductor wafer and chip processing. In semiconductor processing, wafers are modified and transformed in a series of multiple processing steps to create desirable features. Usually, a plurality of identical features are made on each wafer simultaneously by parallel processing to form individual segments on a wafer. Dramatic savings in manufacturing time are achieved by fabricating identical features using parallel or batch processing. In addition, batch processing yields high chip uniformity, and by using certain photolithography and etching methods, very small (sub-micron) features can be precisely fabricated. Accordingly, structures with high feature densities can be fabricated on a very small chip. After processing is completed, the individual segments are cut from wafers in a process known as singulation, to obtain a multiplicity of chips. [Peter Van Zant, “Microchip Fabrication”, 3rd edition, McGraw-Hill 1998].
Semiconductor wafers containing different functional chips can be combined either in final packaging processes by interconnecting different chips or simply by bonding two wafers with different functional chips, then cleaving the stack of wafers. The high efficiency of the semiconductor fabrication process has significantly contributed to the rapid growth of the industry. Highly sophisticated systems have been developed for chip production, packaging, and quality control.
Biochips are arrays of different biomolecules (“probes”) capable of binding to specific targets which are bound to a solid support. There have been essentially two methods to prepare biochips.
The first method involves placing aliquots of solutions containing pre-synthesized probe molecules of interest on a planar substrate, followed by immobilizing the probe molecules in designated positions. For example, probe solutions can be dispensed (“spotted”) on a substrate to form a positionally encoded one-dimensional [Kricka, Larry J., “Immunoassay”, Chapter 18, pages 389-404, Academic Press, 1996] or two-dimensional [U.S. Pat. Nos. 5,807,755 and 5,837,551] probe arrays of customized composition. Molecular probes may be directly attached to a substrate surface or may be attached to solid phase carriers, which in turn are deposited on, or attached to a substrate to form an array. Microparticles (“beads”) represent one type of such carrier. Beads offer the advantage of separating the process of preparing and testing substrates from the process of preparing, applying and testing probe and assay chemistries [U.S. Pat. No. 6,251,691]. Beads of various sizes and compositions have been extensively used in chemical and biochemical analysis as well as in combinatorial synthesis.
The deposition, printing and spotting methods for probe array production have several undesirable characteristics. First, even state-of-the-art deposition and printing technologies only produce arrays of low feature density, reflecting typical spot dimensions of 100 microns and spot-to-spot separations of 300 microns. Second, methods of probe deposition described to date have failed to produce uniform spots, with significant spot-to-spot variations. Third, spotting methods, including such variants as electrophoretic deposition to patterned electrodes [U.S. Pat. No. 5,605,662], require substantial instrumental and logistical support to implement the production of arrays on any significant scale. In particular, spotting methods do not support batch fabrication of probe arrays. That is, while a batch processing format may be used to produce substrates efficiently, the subsequent step of “bio-functionalizing” these substrates by applying chemical or biochemical probes is inefficient, because it does not conform to a batch format but instead requires many individual spotting steps. Thus, this process of manufacturing large numbers of identical functionalized chips is far more time-consuming and expensive than a process that uses parallel processing procedures.
The second method of preparing probe arrays involves in-situ photochemical synthesis of linear probe molecules such as oligonucleotides and peptides using a process similar to photolithography, a standard component of semiconductor processing. These methods have been most widely used in recent years to synthesize, in a parallel set of multi-step photochemical reactions, sets of oligonucleotides in designated sections of glass or similar substrates [U.S. Pat. No. 5,143,854; Proc. Nat. Acad. Sci. USA, 1996, 93: 13555-13560].
Although parallel processing to generate simultaneously a multitude of probe arrays directly on a wafer has the advantage of the scalability and intrinsic improvement in uniformity afforded by batch processing, serious drawbacks exist for the fabrication of probe arrays. First, only simple, relatively short linear molecules are suitably synthesized in a series of single step reactions, and in practice, only arrays of short oligonucleotides have been prepared by this method. Second, the reactions often do not proceed to completion, leading to significant compositional heterogeneity. Third, all semiconductor processing must be completed prior to the introduction of biomolecules, because biomolecules may not be compatible with the harsh environments in certain semiconductor processing steps. This limitation can preclude one from taking full advantage of the wide variety of semiconductor fabrication techniques. Fourth, if functionalization is performed in a batch fabrication format, that fabrication process defines the chemical or biochemical composition (“content”) of each chip on the wafer. That is, to introduce a change in probe design requires that the entire fabrication process be changed accordingly. Customization, while theoretically feasible, requires a change in the sequence of requisite masking steps required for photochemical synthesis of a desired set of probe molecules. The cost and time delays associated with this process renders customization infeasible in practice.